Re: negate if condition - HP UX

This is a discussion on Re: negate if condition - HP UX ; Stephane CHAZELAS wrote: >> [ if ! ... ] > Though /bin/sh on HPUX 11.11 (based on ksh88f) > supports it, /bin/ksh doesn't, but I couldn't find any version > information on it (except that it's some sort of ksh88). ...

+ Reply to Thread
Results 1 to 4 of 4

Thread: Re: negate if condition

  1. Re: negate if condition

    Stephane CHAZELAS wrote:

    >> [ if ! ... ]

    > Though /bin/sh on HPUX 11.11 (based on ksh88f)
    > supports it, /bin/ksh doesn't, but I couldn't find any version
    > information on it (except that it's some sort of ksh88).


    I always believed, that it's actually an "original" ksh88;
    although repeatedly modified by HP with patches (probably
    selected bugfixes).

    If so, this was perhaps done for similar reasons like
    not upgrading the traditional Bourne shell from SVR2 to
    later variants, that is, not changing release-special
    behaviour?

    [xpost c.s.h.hpux, fup2 c.u.shell]

  2. ksh88c on HP-UX? (was: negate if condition)

    Sven Mascheck wrote:
    > Stephane CHAZELAS wrote:


    >> [/bin/ksh] I couldn't find any version information on it
    >> (except that it's some sort of ksh88).


    > I always believed, that it's actually an "original" ksh88;
    > although repeatedly modified by HP with patches (probably
    > selected bugfixes).


    It seems /bin/ksh on HP-UX is a ksh88-c.

    It's missing (at least some of) the fixes from ksh88-d:

    $ a=b set # no output
    $ type true # no plain builtin
    true is an exported alias for :
    $ ksh -c 'echo $0,$1' a b # instead of a,b:
    ksh,a

    but comes with (at least some of) the fixes for ksh88-c and earlier:

    $ set -o emacs; trap : 1 ; (sleep 2; kill -1 $$) & read a 2>/dev/null
    # dumped core before ksh88-c
    $ echo ${x-"{a}"b} # expands to {a}b when x is unset # fixed in ksh88b
    $ $(name=value) ; foobar() ; # core dump for both # fixed in ksh88a
    $ set -A # produces error message in ksh88a
    $ echo $OPTIND # was 0 before ksh88a
    1

    I tested against the list in .

  3. ksh88c on HP-UX? (was: negate if condition)

    Sven Mascheck wrote:
    > Stephane CHAZELAS wrote:


    >> [/bin/ksh] I couldn't find any version information on it
    >> (except that it's some sort of ksh88).


    > I always believed, that it's actually an "original" ksh88;
    > although repeatedly modified by HP with patches (probably
    > selected bugfixes).


    It seems /bin/ksh on HP-UX is a ksh88-c.

    It's missing (at least some of) the fixes from ksh88-d:

    $ a=b set # no output
    $ type true # no plain builtin
    true is an exported alias for :
    $ ksh -c 'echo $0,$1' a b # instead of a,b:
    ksh,a

    but comes with (at least some of) the fixes for ksh88-c and earlier:

    $ set -o emacs; trap : 1 ; (sleep 2; kill -1 $$) & read a 2>/dev/null
    # dumped core before ksh88-c
    $ echo ${x-"{a}"b} # expands to {a}b when x is unset # fixed in ksh88b
    $ $(name=value) ; foobar() ; # core dump for both # fixed in ksh88a
    $ set -A # produces error message in ksh88a
    $ echo $OPTIND # was 0 before ksh88a
    1

    I tested against the list in .

    BTW: It knows the undocumented __ aliases, which Brian Hiles just mentioned
    in <1135049047.271643.166280@f14g2000cwb.googlegroups. com> here (and which
    made me looking at this at all).

  4. Re: ksh88c on HP-UX?

    Sven Mascheck wrote:

    > [...] fixes from ksh88-d:


    Pardon: later fixes also got in here. "a=b set" was valid, though.

+ Reply to Thread